Data science is a vast topic involving different expert profiles. Nowadays, most companies have a data science team, or use a third-party company, because it is a key element for their profitability and their market share. Given the interest in data science, infrastructure, security, and data privacy are crucial elements.
Cloud providers are accelerators for data science & AI projects. However, the Schrems II ruling has made it more complicated to use public cloud with personal data. Fortunately, mitigation measures exist: one of them, pseudonymization, has gained momentum, as many consider it to be the most viable “supplementary measure” for transferring personal data to third countries that do not offer an equivalent level of protection.
In this session, I will present a concrete use case using machine learning models trained with pseudonymized data and discuss the challenges the principles of pseudonymization bring.